Most patients with alcohol use disorders (AUD) never receive formal alcohol treatment, even those already engaged in primary care. Medical management of AUD in primary care settings, including repeated medically-focused brief interventions (BI), has proven effective for decreasing drinking. Monitoring abnormal laboratory tests may engage patients not initially interested in changing drinking, and medications can further improve outcomes among patients with alcohol dependence. However, implementing these evidence-based treatments will not occur by provider education alone-new systems of care delivery are required. The proposed randomized controlled trial tests a Collaborative Care intervention for delivering evidence-based care to AUD patients who do not respond to alcohol screening and BI. The VA is an optimal setting for an initial test of Collaborative Care for AUD because the VA has high rates of annual alcohol screening and clinical informatics systems required for Collaborative Care. The Primary Aims of the proposed trial are to determine whether primary care patients at high risk for current AUD who are offered Collaborative Care, compared to those randomized to usual care, (1) have fewer heavy drinking days and (2) are more likely to be abstinent or drinking below recommended limits without alcohol-related problems at 12 months follow-up. Methods. This randomized controlled trial will enroll 300 adult VA primary care patients < 65 years old (~200 men and 100 women) who are at high risk for current AUD based on their frequency of heavy drinking (> 5 drinks in a day for men and > 4 for women). Patients will be randomized to:1) being offered Collaborative Care for AUD over 12 months or 2) usual care. The Collaborative Care Intervention consists of repeated scheduled visits for BI with a nurse care manager, as well as lab monitoring or medications for AUD when appropriate. The nurse care managers will be supported by a nurse practitioner and an interdisciplinary Collaborative Care Team. Main study outcomes include: 1) the number of heavy drinking days based on the Time Line Follow Back (TLFB), and 2) the proportion of patients who are abstinent or drinking below NIAAA recommended limits without alcohol-related problems based on the TLFB and Short Inventory of Problems (SIP), at 12 months. Generalized estimating equations, adjusted for appropriate covariates, will compare groups based on intention to treat. Secondary outcomes include process measures of treatment engagement, secondary drinking measures, resolution of abnormal lab markers, and inpatient health care utilization. Impact. Broad agreement exists that new approaches for delivering care to patients with AUD are needed, including offering evidence-based care within medical settings. This study, proposed by a New NIH investigator and a team of experts in collaborative care and addictions treatment, tests a system of care delivery proven effective for other mental health conditions. The study team has extensive success designing, testing, and implementing systems for alcohol screening and BI nationwide in VA and conducting clinical trials for AUD, which will inform the proposed study.